Superior aluminium (Al) tolerance of Stylosanthes is achieved mainly by malate synthesis through an Al-enhanced malic enzyme, SgME1


  • Lili Sun,

    1. State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Root Biology Center, South China Agricultural University, Guangdong, China
    2. Institute of Tropical Crop Genetic Resources, Chinese Academy of Tropical Agriculture Sciences, College of Agriculture, Hainan University, Hainan, China
    Search for more papers by this author
  • Cuiyue Liang,

    1. State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Root Biology Center, South China Agricultural University, Guangdong, China
    Search for more papers by this author
  • Zhijian Chen,

    1. State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Root Biology Center, South China Agricultural University, Guangdong, China
    Search for more papers by this author
  • Pandao Liu,

    1. State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Root Biology Center, South China Agricultural University, Guangdong, China
    2. Institute of Tropical Crop Genetic Resources, Chinese Academy of Tropical Agriculture Sciences, College of Agriculture, Hainan University, Hainan, China
    Search for more papers by this author
  • Jiang Tian,

    Corresponding author
    1. State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Root Biology Center, South China Agricultural University, Guangdong, China
    Search for more papers by this author
  • Guodao Liu,

    Corresponding author
    1. Institute of Tropical Crop Genetic Resources, Chinese Academy of Tropical Agriculture Sciences, College of Agriculture, Hainan University, Hainan, China
    Search for more papers by this author
  • Hong Liao

    1. State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Root Biology Center, South China Agricultural University, Guangdong, China
    Search for more papers by this author


  • Stylosanthes (stylo) is a dominant leguminous forage in the tropics. Previous studies suggest that stylo has great potential for aluminium (Al) tolerance, but little is known about the underlying mechanism.
  • A novel malic enzyme, SgME1, was identified from the Al-tolerant genotype TPRC2001-1 after 72 h Al exposure by two-dimensional electrophoresis, and the encoding gene was cloned and characterized via heterologous expression in yeast, Arabidopsis thaliana and bean (Phaseolus vulgaris) hairy roots.
  • Internal Al detoxification might be mainly responsible for the 72 h Al tolerance of TPRC2001-1, as indicated by 5.8-fold higher root malate concentrations and approximately two-fold higher Al concentrations in roots and root symplasts of TPRC2001-1 than those of the Al-sensitive genotype Fine-stem. An accompanying increase in malate secretion might also reduce a fraction of Al uptake in TPRC2001-1. Gene and protein expression of SgME1 was only enhanced in TPRC2001-1 after 72 h Al exposure. Overexpressing SgME1 enhanced malate synthesis and rescued yeast, A. thaliana and bean hairy roots from Al toxicity via increasing intracellular malate concentrations and/or accompanied malate exudation.
  • These results provide strong evidence that superior Al tolerance of stylo is mainly conferred by Al-enhanced malate synthesis, functionally controlled by SgME1.


Aluminum (Al) is a major constraint limiting crop yield on acidic soils, which occupy approximately 50% of the arable land worldwide (Kochian et al., 2004). In the last few decades, great efforts have been made to understand the mechanisms of Al tolerance, including external and internal Al tolerance (Kochian et al., 2005; Horst et al., 2010). External mechanisms inhibit Al uptake by roots and reduce detrimental interactions between Al and sensitive sites in the apoplast (Delhaize et al., 2012). Examples include organic acid exudation (Hoekenga et al., 2006; Magalhaes et al., 2007), root mucilage secretion (Horst et al., 1982), cell wall modification (Horst et al., 1999; Huang et al., 2010; Zhu et al., 2012) and rhizosphere pH modification (Pellet et al., 1997). Among these, the exclusion of Al from the root apex by the efflux of Al-activated organic acids is well supported by strong genetic and molecular evidence in many crop species (Li et al., 2000; Ma et al., 2001; Sasaki et al., 2004; Magalhaes et al., 2007; Piñeros et al., 2008; Liang et al., 2013; Liu et al., 2013).

Despite the exclusion of Al via external mechanisms, some phytotoxic Al inevitably enters the root symplasts, especially after long-term Al exposure (Lazof et al., 1994). Therefore, it would be reasonable to hypothesize that plant species grown in acidic soils would evolve internal Al detoxification. This is supported by observations of Metali et al. (2012) in which internal tolerance mechanisms are common in tropical and subtropical plants where acidic soils are endemic. Internal tolerance mechanisms typically detoxify Al once it enters the cell through means such as forming harmless complexes with organic ligands (Ma et al., 1997; Watanabe et al., 1998), or sequestering Al to organelles (de Andrade et al., 2011; Huang et al., 2012). However, in most crop species, little is known about the mechanisms underlying internal Al tolerance.

Rice has been widely demonstrated as one of the most Al-tolerant crop species with both internal and external mechanisms involved (Yang et al., 2008; Yamaji et al., 2009; Xia et al., 2010; Yokosho et al., 2011; Huang et al., 2012). This tolerance has been associated with multiple pathways regulated in large part by OsART1, a member of the Cys2His2-type zinc-finger transcription factor family (Yamaji et al., 2009). The Al-activated citrate efflux pathway controlled by the MATE transporter, OsFRDL4, has proven to be an important external mechanism (Yokosho et al., 2011). Internally, OsALS1, a half-size ABC transporter located in the tonoplast, is a critical internal Al detoxification player in rice (Huang et al., 2012), which may sequester Al taken up by Nrat1, a plasma membrane-localized Al transporter (Xia et al., 2010), or other pathways into the vacuoles.

In some crops, internal Al may be chelated by organic acids or other compounds intracellularly. Malate, citrate and oxalate are the most common organic acids involved in this internal Al detoxification. For example, oxalate was the predominant intracellular ligand in buckwheat (Fagopyrum esculentum) leaves (Ma et al., 1998; Ma & Hiradate, 2000), while hydrangea (Hydrangea macrophylla) leaves strongly chelated Al with citrate (Ma et al., 1997). Overexpressing genes associated with organic acid synthesis in plants may result in increased organic acid efflux and thus greatly enhance Al tolerance (de la Fuente et al., 1997; Tesfaye et al., 2001; Han et al., 2009; Wang et al., 2010), which indicates a dependency of increasing organic acid efflux on internal organic acid pools. However, specifics of the roles of internal organic acid pools playing in Al tolerance remain unclear.

Stylo (Stylosanthes) is a dominant leguminous forage crop in the tropics and subtropics. Approximately 1 Mha of planting area is occupied by stylo in Queensland (Noble et al., 2000), which is one of the largest rangelands in Australia. In China, there were at least 113 Kha stylo planted in Guangdong province in 1992 (Devendra & Seré, 1993). Stylo originates from the tropics, and is well adapted to tropical and subtropical areas (Noble et al., 2000) where soils are often highly acidic. After a hundred years of cultivation in the tropics, stylo plants have developed strong adaptability to acidic soils, which might involve Al tolerance (Andrew et al., 1973; de Carvalho et al., 1980). Yet, to date, research on mechanisms of stylo responses and tolerance to Al is very limited. In this study, detailed physiological and molecular studies on stylo Al tolerance were conducted. Superior Al tolerance in stylo was found to be mainly conferred by enhancement of malate concentrations in roots and accompanied slightly by malate exudation from roots. Furthermore, this malate synthesis in roots is demonstrated to be controlled by a novel malic enzyme (ME), SgME1.

Materials and Methods

Plant growth conditions

Two stylo (Stylosanthes guianensis Aubl.) genotypes contrasting in adaptation to acidic soils were selected based on the field experiment previously reported by Du et al. (2009). Between them, TPRC2001-1 exhibited phosphorous (P) efficiency and Al tolerance that was superior to those traits in Fine-stem.

For Al tolerance characterization, rice (Oryza sativa L.) cv Xiangnuo 1 (XN1) – previously identified as an Al-tolerant genotype (Xu et al., 2004) – was used as a standard for evaluating Al tolerance of the two tested stylo genotypes. Uniform seeds of rice and stylo were germinated on solid Murashige and Skoog (MS) medium for 36 h, and then germinated seeds with emerging radicals (0.5–1 cm in length) were transplanted to boxes containing 0.5 mM CaCl2 (pH 4.2) with or without 100 μM AlCl3. Roots were harvested at 0, 24 and 72 h after initiating Al treatment. Root length was measured by ImageJ software (National Institutes of Health, Bethesda, MD, USA) using pictures scanned with an Epson 1640× L scanner (GE Healthcare, Piscataway, NJ, USA).

In physiological and molecular experiments, stylo seedlings were precultured in nutrient solution as previously described (Liao et al., 2006) for 1 month, and then exposed to 0.5 mM CaCl2 (pH 4.2) containing 0 or 100 μM AlCl3 for an additional 24 or 72 h. After collecting samples for measuring secreted organic acids, roots and shoots were separately harvested for determining the content of organic acids and Al, as well as for extracting RNA and protein. Al content was measured with an inductively coupled plasma atomic emission spectrometer 710-ES (Varian, Palo Alto, CA, USA) after ash digestion as described by Murphy & Riley (1962). Symplasts were separated for Al detection according to the description of Wang et al. (2004). Each treatment in all the experiments had four biological replicates.

Detection of internal and external organic acids

For collecting secreted organic acids, seedlings were transferred to centrifuge tubes containing 20 ml 0.5 mM CaCl2 collecting solution (pH 4.2) with or without 100 μM AlCl3 for 6 h. The samples were stored at −80°C, concentrated to dry powder using a freeze drying vacuum system (Labconco, Kansas, MO, USA), and then dissolved in 1 ml Millipore water before taking measurements. For the extraction of internal organic acids, 0.3-g root samples were first homogenized in 1.2 ml 0.25 M HCl, followed by heating at 80°C for 20 min with intermittent shaking, and then centrifuged at 12 000 g for 10 min. Finally, the supernatant containing organic acids was passed through a filter (0.45 μm) and analyzed with a HPLC 1260 Infinity LC (Agilent, USA) following the method described by Xu et al. (2006). To identify and determine the amount of organic acids, the retention time and absorption spectra in samples were compared with known standards.

Identification of Al-responsive proteins via two-dimensional (2D) electrophoresis

Total root protein was extracted according to modified procedures (Chen et al., 2011). Protein concentrations were determined according to the 2-D Quant Kit manual (GE Healthcare). An Ettan IPGphor3 apparatus (GE Healthcare) was used for isoelectric focusing (IEF) with immobilized pH gradient (IPG) strips (pH 4.0–7.0 linear gradient, 24 cm). For each treatment, 2D electrophoresis was repeated three times. Differentially expressed protein spots were digested and MALDI-TOF/TOF MS analyzed using an ABI 4700 TOF-TOF (Applied Biosystems, USA). The data were searched by GPS Explorer (v3.6) with the search engine MASCOT (v2.1) against the National Center for Biotechnology Information database (version NBCInr 20100724).

Isolation and characterization of SgME1

Total root RNA was extracted according to the manual for the TRIzol reagent (Invitrogen) with slight modifications. Briefly, 0.2 g root tissue powder was mixed with 1.5 ml TRIzol reagent and a pinch of PVPP and vortexed thoroughly. The homogenate was then incubated at 60°C for 30 min followed by incubation on ice for 10 min. After centrifuging at 12 000 g for 10 min at 4°C, the supernatant was mixed with 0.3 ml chloroform and incubated on ice for 10 min, and, centrifuged again at full speed for 15 min at 4°C. The aqueous phase was then mixed with precipitation buffer (0.4 ml isopropanol containing 1.2 M NaCl and 0.8 M sodium citrate) and incubated at −20°C for 8 h. Pellets were retrieved by centrifugation at 12 000 g for 10 min at 4°C. After washing with 75% ethanol, RNA was air dried and dissolved with DEPC treated water. First strand cDNA was synthesized from 2 μg total RNA using M-MLV reversed transcriptase (Promega) according to the manual.

In order to clone the SgME1 gene, primers were designed according to the conserved motif sequences of four ME genes of Phaseolus vulgaris (J03825.1), Glycine max (XM_003526456.1), Vitis vinifera (XM_003631725.1) and Medicago truncatula (XM_003630679.1). A 443-bp fragment was amplified by using SgME1-EST-F and SgME1-EST-R (Supporting Information Table S1). The PCR product was then cloned into pGEM-T vector (Promega) and sequenced. Rapid amplification of cDNA ends (RACE) was performed to amplify the 5′- and 3′-ends of the cDNA with two specific primers – SgME1-RACE5′-R and SgME1-RACE3′-F (Table S1). The sequences of the three fragments were analyzed and combined together in the MEGA v4.1 program to generate a full-length cDNA of SgME1. The sequence of SgME1 was deposited in the NCBI database (accession no. JX164253). A phylogenetic tree was constructed with MEGA 4.1.

The subcellular localization of SgME1 was investigated by transforming onion (Allium cepa) epidermal cells with the plasmids of the SgME1-GFP using the primers of SgME1-GFP-F and SgME1-GFP-R (Table S1) fusion protein construct and a 35S:GFP empty vector by particle bombardment (PDS/1000, Bio-Rad). Propidium iodide staining was used for distinguishing the cell wall. Plasmolysis was conducted by treating the onion cells with 30% sucrose solution.

Production of SgME1 antiserum and protein immunoblot analysis

SgME1 antiserum was produced by synthesizing polypeptides according to the SgME1 protein sequence (Abmart, Shanghai, China). Western blotting was carried out as previously described (Liang et al., 2010). Briefly, the total protein in TPRC2001-1 roots was extracted and separated by 2D electrophoresis, then electrophoretically transferred to polyvinylidene difluoride membranes (Bio-Rad) and hybridized with SgME1 antibody (1 : 200 dilution). Blots were developed by the reaction with an alkaline phosphatase-tagged secondary antibody. Additionally, the total protein of TPRC2001-1 roots, transgenic Arabidopsis thaliana plants and bean hairy roots were separately run through sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) for Western blotting analysis.

SgME1 expression analysis via quantitative real-time PCR

Quantitative real-time PCR (qPCR) was conducted using a Rotor-Gene 3000 qPCR system (Corbett Research, Mortlake, Victoria, Australia) and SYBR Premix Ex Taq II (Takara, Shiga, Japan). Primers SgME1-RT-F and SgME1-RT-R (Table S1) were used to assay for SgME1 expression. The housekeeping genes SgEF-1a (accession no. JX164254) with primers SgEF-1a-F and SgEF-1a-R, AtEF-1a (accession no. NM_100666) with primers AtEF-1a-F and AtEF-1a-R, and PvEF-1a (PvTC3216 from the Dana-Farber Cancer Institute Computational Biology and Functional Genomics Laboratory) with primers PvEF-1a-F and PvEF-1a-R were used as internal controls for stylo, A. thaliana and common bean, respectively (Table S1). All of the gene expression analysis had three biological replicates. Relative expression levels were calculated from the ratio of expression levels of SgME1 to expression levels of the housekeeping gene.

Yeast transformation

The entire open reading frame (ORF) of SgME1 was amplified by PCR using the primers SgME1-pYES2-F with a BamHI site and SgME1-pYES2-R containing an EcoRI site (Table S1), and cloned into the pGEM-T vector. After sequence confirmation, the SgME1 fragment was released through BamH1 and EcoR1 digestion and cloned into a pYES2 vector (Invitrogen) digested with the same enzymes. The SgME1-pYES2 or the empty vector pYES2 was separately introduced into yeast strain INVSC1 using an SC easy comp transformation kit (Invitrogen). Transformants were selected on uracil-deficient medium (SC-uracil medium) containing 2% glucose, 0.67% yeast nitrogen base without amino acids, 0.1% uracil dropout mix and 2% agar. Three independent yeast colonies harboring SgME1-pYES2 constructs and one empty vector transformant were selected and grown in liquid SC-uracil medium. The OD600 value of the precultured yeast was adjusted from 0.6 to 0.2. The induction plate, which contained 2% galactose, 1% raffinose, 0.67% yeast nitrogen base without amino acids, 0.1% uracil dropout mix, and 2% agar, was surfaced-plated with or without 4 mM AlCl3. After spotting at four yeast-cell dilutions (optical densities at 600 nm of 0.2, 0.02, 0.002 and 0.0002), plates were incubated for 2 d or 4 d at 30°C.

For determination of malate accumulation in yeast cells, colonies carrying an empty vector or a SgME1 vector were grown in 100 ml liquid induction medium for 1 d. Then the yeast cells were transferred to centrifuge tubes, pelleted by centrifugation at 6000 g for 10 min at 4°C and washed twice with 1 ml Millipore water. The pellets were freeze dried in a lyophilizer (Labconco), and samples were then ground in liquid nitrogen, dissolved in 1.2 ml 0.25 M HCl, heated at 80°C for 20 min and centrifuged at 12 000 g for 10 min. Finally, malate in the supernatant was detected as described above.

Heterologous expression of SgME1 in A. thaliana

The ORF of SgME1 was amplified using primers SgME1-pYL-F and SgME1-pYL-R (Table S1) and cloned into the pYLRNAi binary vector. The construct was transformed into the Agrobacterium tumefaciens strain GV3101, which was then used for A. thaliana transformation using the floral dip method (Clough & Bent, 1998). Gene expression of SgME1 and malate concentrations in transgenic plants were analyzed as described above. Two representative transgenic lines were employed for Al tolerance assayed through measuring relative root elongation. After A. thaliana seeds were surface-sterilized and germinated on solid MS medium for 5 d, uniform seedlings were grown on solid agar medium supplied with 4.3 mM CaCl2 and 3% sucrose, as well as 0 or 400 μM AlCl3 at pH 4.2 for 2 d, respectively. Relative root elongation was calculated for roots grown with Al added relative to controls without Al addition (Xia et al., 2010). In addition, uniform seedlings were precultured in hydroponic MS medium with a shaking speed of 180 rpm for 10 d, and then were washed with sterilized water several times. The precultured seedlings were incubated in solution containing 0.5 mM CaCl2 (pH 4.2) with or without 100 μM AlCl3. After 24 h of Al treatment, the incubated solution was sampled and organic acids were extracted from plants for identification and quantification.

SgME1 overexpression in common bean hairy roots

The pYLRNAi vector containing the SgME1 coding region was introduced into bean (Phaseolus vulgaris) hairy roots by Agrobacterium rhizogenes strain K599 mediated transformation (Liang et al., 2010). The gene expressions and malate concentrations of transgenic lines were verified by qPCR and HPLC analysis in comparison to an empty vector control, respectively. Two independent overexpressing lines and one control line were selected to evaluate the effects of Al toxicity on hairy roots. Transgenic hairy roots were grown on solid agar medium supplied with 4.3 mM CaCl2 and 3% sucrose, as well as 0 or 1.2 mM AlCl3 at pH 4.2 for 2 d, respectively. The relative root elongation was calculated after Al treatment as described above. Subsequently, transgenic roots were shaken slightly in solution consisting of 0.5 mM CaCl2 with or without 100 μM AlCl3 for 1 d. The fluid was then collected for detection of organic acids exuded from hairy roots as before. A portion of the hairy root tips were examined under scanning electron microscopy to visualize morphological changes of the root apex (Liao et al., 2006), while the rest had organic acids extracted and assayed, or were digested for Al detection as described above.

Statistical analysis

All the statistical analysis was performed by two-way ANOVA, Tukey's test and t-test using the SPSS program (v13.0; SPSS Institute, Chicago, IL, USA).


Superior Al tolerance of TPRC2001-1

The two stylo genotypes significantly differed in their tolerance to Al toxicity as indicated by root growth in 100 μM Al (Fig. 1). This was consistent with previous observations in a field experiment on acidic soils (Du et al., 2009). After 72 h Al exposure, root growth of both TPRC2001-1 and Fine-stem was reduced, while in TPRC2001-1 it was reduced less than that in Fine-stems, which led to higher root growth in TPRC2001-1 than that in Fine-stem (Fig. 1b). Interestingly, there was no genotypic difference in root growth between the two stylo genotypes after 24 h Al treatment (Fig. 1a). This suggests that the superior Al tolerance of TPRC2001-1 mainly involves 72 h Al exposure responses.

Figure 1.

Root growth of stylo and rice plants after 24 h and 72 h aluminium (Al) exposure. Germinated seedlings were exposed to 100 μM AlCl3 for 24 h (a) and 72 h (b). XN1, a rice (Oryza sativa) genotype (hatched bars); TPRC2001-1 (white bars) and Fine-stem (gray bars), two stylo (Stylosanthes guianensis) genotypes. Each bar represents the mean of four replicates ± SE. To obtain homogeneity of variances, the data were logarithmically transformed before analysis. F-values of ANOVAs: (a) 23.6 for genotype (< 0.001), 92.53 for Al treatment < 0.001), 6.76 for interaction between treatment and genotype (< 0.05); (b) 51.63 for genotype (< 0.001), 106.26 for Al treatment < 0.001), 16.36 for interaction between treatment and genotype (< 0.001). Different letters indicate significant differences at 0.05 levels.

In order to determine how superior the Al tolerance of TPRC2001-1 is, the root growth responses of TPRC2001-1 and a highly Al-resistant rice variety XN1 were compared. XN1 had significantly higher root growth than both stylo genotypes at 24 h after Al treatment (Fig. 1a). However, the root growth of TPRC2001-1 and XN1 was similar after 72 h of Al treatment (Fig. 1b), indicating that Al tolerance in TPRC2001-1 is comparable to that in XN1 after 72 h Al exposure.

Al concentrations in stylo roots were also investigated after 24 and 72 h of 100-μM Al treatment. Al concentrations in the roots of both genotypes were extremely low without Al application (data not shown). However, a significant increase of Al absorption was observed after Al treatment (Fig. 2a). Al concentrations in roots of TPRC2001-1 were approximately two-fold higher than those in Fine-stem after 24 and 72 h of Al treatment (Fig. 2a). Furthermore, Al concentrations in the root symplasts of TPRC2001-1 were around twice those in Fine-stem at both 24 and 72 h of Al exposure (Fig. 2b). These results strongly suggest that internal detoxification might be a major Al tolerance mechanism in TPRC2001-1.

Figure 2.

Aluminium (Al) concentrations in roots and root symplasts of two stylo (Stylosanthes guianensis) genotypes after 24 and 72 h Al treatments. Plants were precultured in nutrient solution for 1 month, and then separately subjected to 100 μM AlCl3 treatment for 24 and 72 h. Al concentrations in the roots (a) and root symplasts (b) were measured after Al treatment. TPRC2001-1, white bars; Fine-stem, gray bars. Each bar represents the mean of five replicates ± SE. F-values of ANOVAs: (a) 3.81 for treatment time (ns, nonsignificant difference at 0.05 levels), 47.91 for genotype (< 0.001), 0.29 for interaction between treatment time and genotypes (ns); (b) 4.33 for treatment time (ns), 31.76 for genotype (< 0.001), 0.34 for interaction between treatment time and genotype (ns). Different letters indicate significant differences at 0.05 levels.

Malate concentrations in both roots and root exudates of TPRC2001-1 increased only after 72 h Al exposure

In order to determine the mechanisms underling stylo Al tolerance, malate concentrations in stylo roots and root exudates were measured. After treatment with 100 μM Al for 24 h, internal malate concentrations in the roots of both genotypes had not changed, and no genotypic difference was detected either (Fig. 3a). Malate exudation from Fine-stem roots was higher than that from TPRC2001-1 roots after 24 h of Al treatment (Fig. 3c).

Figure 3.

Effects of aluminium (Al) toxicity on malate synthesis and exudation in two stylo (Stylosanthes guianensis) genotypes. Malate concentrations in roots (a, b) and root exudates (c, d) were detected after treatments by 0 (−Al) and 100 μM AlCl3 (+Al) for 24 h (a, c) and 72 h (b, d). −Al, white bars; +Al, gray bars. Each bar represents the mean of four replicates ± SE. F-values of ANOVAs: (a) 0.12 for genotype (ns, nonsignificant difference at 0.05 levels), 1.51 for Al treatment (ns), 0.035 for interaction between treatment and genotype (ns); (b) 78.51 for genotype (< 0.001), 18.12 for Al treatment (< 0.01), 58.13 for interaction between treatment and genotype (< 0.001); (c) 8.24 for genotype (< 0.05); 1.90 for Al treatment (ns), 2.22 for interaction between treatment and genotype (ns); (d) 6.09 for genotype (< 0.05), 9.91 for Al treatment (< 0.05), 9.71 for interaction between treatment and genotypes (< 0.05). ns, no significant difference at 0.05 levels. Different letters mean significant difference at 0.05 levels.

After 72 h of Al treatment, the malate concentrations in roots of TPRC2001-1, but not Fine-stem, were significantly enhanced by Al treatment (Fig. 3b). Root malate concentrations were 5.8 times higher in TPRC2001-1 than in Fine-stem (Fig. 3b). Malate exudation from TPRC2001-1 also increased significantly with increasing internal malate concentrations, and was 2.5 times higher than that from Fine-stem after 72 h Al exposure (Fig. 3d). These results imply that greatly enhanced internal malate accumulation in Al-tolerant stylo genotype TPRC2001-1 might be a tolerance mechanism to deal with 72 h internal Al stress in the symplasts, which might spur increased malate exudation to slightly reduce Al uptake.

Isolation and characterization of 72 h Al exposure enhanced malic enzyme, SgME1

In order to investigate potential key proteins in stylo malate accumulation and/or malate exudation for Al detoxification, two dimensional (2D) electrophoresis was conducted to hunt for differentially expressed proteins in TPRC2001-1 roots between the two Al treatments (i.e. 100 μM Al and 0 μM Al treatment). One protein, which only appeared after 72 h Al treatment and was markedly stimulated by Al (Fig. 4), was selected and identified by MALDI TOF/TOF MS analysis. Results showed that the protein is homologous to NADP-dependent malic enzyme (ME).

Figure 4.

Identification of SgME1 through 2D electrophoresis. SgME1 protein accumulation in a stylo (Stylosanthes guianensis) genotype, TPRC2001-1 treated by 0 μM (a, c) and 100 μM AlCl3 (b, d) for 24 h (a, b) and 72 h (c, d). The SgME1 protein spot is boxed.

Based on conserved motif sequences in MEs reported from other plant species, the full-length gene encoding this protein was cloned. Because it is the first reported gene encoding an NADP-dependent ME in stylo, it was designated as SgME1. Western blotting was further performed to confirm whether the ME enzyme identified in the 2D gel is encoded by SgME1. Results showed that the SgME1 antibody could specifically immunoblot the ME protein in a 2D gel (Fig. S1a). In addition, SgME1 protein in Al treated TPRC2001-1 roots, SgME1 overexpression A. thaliana plants and transgenic bean hairy roots was also detected using the SgME1 antibody via Western blotting (Fig. S1b). These results strongly indicate that SgME1 is the gene that encodes SgME1 identified from the 2D electrophoresis.

Phylogenetic analysis showed that the SgME1 protein had high (over 80%) similarity to AtME2 (AtNP196728) in A. thaliana (Fig. S2), which catalyzes pyruvate reductive carboxylation to produce malate (Gerrard Wheeler et al., 2009). When expressed in onion epidermal cells, the fluorescence signal of SgME1-GFP was observed throughout the cytosol (Fig. S3), which corresponds to the localization of AtME2. These results taken together suggest that SgME1 (EC and AtME2 might fulfil a similar function in malate synthesis.

Expression of SgME1 was only enhanced by 72 h Al exposure in roots of the Al-tolerant genotype, TPRC2001-1

In order to determine the expression pattern of SgME1 in both stylo genotypes responsive to Al toxicity, qPCR analysis was further conducted. After 24 h Al exposure, no significant change of SgME1 was found for either genotypes (Fig. 5a). However, after 72 h Al exposure, the expression of SgME1 was significantly upregulated in TPRC2001-1, but not in Fine-stem (Fig. 5b), with expression in TPRC2001-1 rising to twice that in Fine-stem with Al application (Fig. 5b). This indicates that the response of SgME1 to Al treatment corresponds to that of the SgME1 protein accumulation. However, expression levels of SgME1 in roots of both TPRC2001-1 and Fine-stem remained unchanged during the treatment without Al application (Figs 5, S4).

Figure 5.

Expression of SgME1 in response to aluminium (Al) exposure in two stylo (Stylosanthes guianensis) genotypes. Expression of SgME1 in both TPRC2001-1 and Fine-stem treated with 0 (−Al) and 100 μM AlCl3 (+Al) for 24 h (a) and 72 h (b). −Al, white bars; +Al, gray bars. Each bar represents the mean of four replicates ± SE. F-values of ANOVAs: (a) 44.81 for genotype (< 0.001), 1.49 for Al treatment (ns, nonsignificant difference at 0.05 levels), 0.68 for interaction between treatment and genotype (ns); (b) 34.32 for genotype (< 0.001), 13.40 for Al treatment (< 0.01), 9.82 for interaction between treatment and genotype (< 0.05). Different letters indicate significant differences at 0.05 levels.

Overexpression of SgME1 conferred yeast Al tolerance

The function of SgME1 in malate synthesis was first determined via overexpressing the gene in yeast cells. Results showed that yeast cells expressing SgME1 have higher malate accumulation than the empty vector control (Fig. 6a). Furthermore, under normal conditions (−Al), the growth of SgME1-overexpression yeast cells was similar to that of the empty vector control. However, in the presence of 4 mM Al, SgME1-overexpression yeast cells grew much better than those harboring the empty vector control (Fig. 6b). These results imply that higher malate accumulation mediated by overexpressing SgME1 might be attributed to confer Al tolerance in transgenic yeast cells.

Figure 6.

Effects of SgME1 overexpression in yeast. (a) Malate concentrations in yeast cells after incubation in liquid medium for 1 d. (b) Growth of yeast cells in response to aluminium (Al) treatment. Yeast cells (INVSC1) carrying empty vector pYES2 or SgME1 were spotted on SC (pH 4.2) medium with or without 4 mM AlCl3 for 2 and 4 d. CK, empty vector control; OX, transgenic clones carrying SgME1. Each bar represents the mean of four replicates ± SE. Asterisks indicate significant differences between transformants and the empty vector control in t-tests: **, 0.001 < < 0.01; ***, < 0.001.

Overexpression of SgME1 in A. thaliana enhanced Al tolerance

The function of SgME1 was further investigated by overexpressing SgME1 in A. thaliana. Results of qPCR showed that SgME1 was highly expressed in two transgenic lines (i.e. OX1 and OX2) with greatly increased malate concentrations (Fig. S5). The overexpression lines showed greater Al tolerance as indicated by higher relative root elongation compared with the control (Fig. 7a,b). The malate concentrations in the SgME1-overexpression A. thaliana after Al treatment had no significant increase compared to those in wild-type (WT) (Fig. 7c). However, the malate exudation rate was 95% in OX1 and 99% in OX2 higher than that in WT lines, respectively (Fig. 7d). These results further verified that overexpressing SgME1 could confer Al tolerance in A. thaliana.

Figure 7.

Physiological characterization of Arabidopsis thaliana with SgME1 overexpression. Growth (a) and relative root elongation (b) of 5-d-old A. thaliana plants after 2 d exposure to 400 μM aluminium (Al) in solid media. Bars, 1 cm. Malate concentrations (c) and exudation rates (d) in A. thaliana plants detected from 24 h 100 μM Al treatments of 15 d old plants. WT, wild-type; OX, A. thaliana transgenic lines overexpressing SgME1. Each bar represents the mean of four replicates ± SE. Asterisks indicate significant differences between transgenic lines and WT plants in t-tests: *, < 0.05.

Overexpression of SgME1 in bean hairy roots ameliorated Al toxicity

Two overexpressing-SgME1 (OX) lines of transgenic bean hairy roots were generated and verified by qPCR analysis (Fig. S6a). The root malate concentrations were dramatically increased in the SgME1 overexpression lines without Al addition (Fig. S6b). The Al tolerance of the transgenic hairy roots was evaluated after treatment with Al for 48 h. SgME1-overexpression lines showed higher Al tolerance as indicated by better root growth (Fig. 8a,b), as well as less damage of the epidermal cells observed through SEM (Fig. 8c) than in the empty vector control. The relative root length of transgenic lines OX1 and OX2 was 62% and 76%, respectively, higher than that of the control after Al treatment (Fig. 8b). However, as seen in Fig. 8(d), the Al concentrations in the two SgME1 overexpression lines were similar to those in the empty vector control, suggesting that the improved Al tolerance in SgME1-overexpression hairy root lines is not mainly caused by excluding Al fromroots.

Figure 8.

Phenotypic analysis of bean (Phaseolus vulgaris) hairy roots overexpressing SgME1. Transgenic bean hairy roots were subjected to 1.2 mM AlCl3 on plates for 2 d before measurement. (a) Growth of transgenic bean hairy roots. Bars, 5 mm. (b) Relative root elongation. (c) Root tip morphology imaged using scanning electron microscopy. Bars, 200 μm. Aluminium (Al) concentrations (d), malate concentrations (e) and malate exudation rates (f) in transgenic bean hairy roots determined after 100 μM Al treatments for 24 h. CK, control hairy roots with the empty vector; OX1 and OX2, two transgenic lines overexpressing SgME1. Each bar represents the mean of four replicates ± SE. Asterisks indicate significant differences between transgenic lines and empty vector controls in t-tests: *, < 0.05; **, 0.001 < < 0.01.

Furthermore, in comparison to those in the control line with Al application, the malate concentrations were 41% in OX1 and 29% in OX2 more, while the malate exudation rates were 2.2-fold in OX1 and 1.4-fold in OX2 higher, respectively (Fig. 8e,f). The fact that Al concentrations (Fig. 8d) in roots of the two SgME1 overexpressing lines did not decrease even when the malate exudation was much higher than those of the control, implies the great ability of internal Al detoxification through enhancing malate accumulation in bean SgME1-overexpression hairy roots.


In tropical and subtropical areas, Al toxicity is one of the most limiting factors for crop productivity (von Uexküll & Mutert, 1995; Kochian et al., 2004). Plants have developed special mechanisms to detoxify Al internally and externally. Recently, Metali et al. (2012) analyzed foliar Al concentrations of over 1000 plant species, and found that plants from tropical regions accumulate more Al in their leaves than plants from other regions, suggesting that internal Al detoxification might be more pronounced in tropical plants. Consistently, it has been suggested that Al accumulation in the cell walls of the leaf epidermal cells is important to Al tolerance in tea (Camellia sinensis) (Tolrá et al., 2011). However, the underlying mechanisms of internal Al detoxification in most tropical plants are still unclear. In this study, the Al tolerance of Stylosanthes was evaluated, and then the physiological and molecular mechanisms underlying its Al tolerance were elucidated through proteomic and transgenic approaches.

First, genotypic variations for Al tolerance in stylo were found, and a tolerant genotype was identified. This genotype, TPRC2001-1, has comparable Al tolerance to the Al-tolerant rice genotype XN1, as indicated by higher root growth than that in Fine-stem after 72 h Al exposure (Fig. 1b). Because rice has been characterized as one of the most Al-tolerant crops (Foy, 1988), and XN1 has been used in several studies as the Al-tolerant rice genotype (Xu et al., 2004; Yang et al., 2007; Zhang et al., 2007), TPRC2001-1 is indicated to have superior Al tolerance. Through analysis of Al accumulation in stylo genotypes contrasting in Al tolerance, we found that the Al-tolerant genotype TPRC2001-1 had much higher Al concentrations in both total roots and root symplasts than the Fine-stem genotype under both 24 and 72 h Al treatments (Fig. 2). This suggests that the superior Al tolerance of TPRC2001-1 is due mainly to internal Al detoxification rather than external Al detoxification as in most tropical plant species (Metali et al., 2012).

Using 2D electrophoresis, a novel NADP-dependent malic enzyme (SgME1, EC was identified, exhibiting high sequence similarity and similar subcellular localization of AtME2 from A. thaliana (Figs S2, S3) (Gerrard Wheeler et al., 2009). Furthermore, overexpressing SgME1 in yeast, A. thaliana and common bean hairy roots resulted in significant increase of malate concentrations (Figs 6a, S5b, S6b), which further verified that the SgME1 identified in this study is a novel malic enzyme, functioning in malate synthesis. Furthermore, the expression of SgME1 was only enhanced in the roots of TPRC2001-1 but not in Fine-stem (Fig. 5b). The malate concentrations also only increased in TPRC2001-1 roots after 72 h Al treatment (Fig. 3b). Taken together, these results imply that increased expression of SgME1 and enhanced protein accumulation of SgME1 are involved in controlling Al-mediated malate synthesis in the Al-tolerant stylo genotype, TPRC2001-1. In addition, even though malate exudation from TPRC2001-1 roots after 72 h Al treatment increased three-fold relative to Al-free controls (Fig. 3d), we infer that the increased malate exudation might be due to the increase of internal malate in roots. Similar results have been reported by Wang et al. (2010) in tobacco and by Tesfaye et al. (2001) in alfalfa. Overexpression of malate dehydrogenase not only increases the internal malate concentrations in roots, but also enhances malate efflux (Tesfaye et al., 2001; Wang et al., 2010).

Although a transformation system in stylo has not yet been established, the function of SgME1 on Al tolerance was verified by heterologous overexpression of SgME1 in yeast cells, A. thaliana and transgenic bean hairy roots. We found that overexpression of SgME1 confers Al resistance in yeast cells, A. thaliana, as well as transgenic bean hairy roots (Figs 6-8), providing strong evidence that the expression of SgME1 is significantly correlated with Al tolerance in yeast and planta. The conferred Al tolerance of transformants might be the result of enhanced malate synthesis due to SgME1 overexpression (Figs 6a, S5b, S6b). Increased malate synthesis in transgenic lines could directly increase internal malate concentrations, and also subsequently increase malate exudation.

However, we found that the portion of malate in roots for exudation under Al stress is largely plant species-dependent, which could be supported by our results using the transgenic approach. In A. thaliana, most of the malate in roots of overexpressing SgME1 lines were secreted when treated with Al, and resulted in no difference of malate concentrations in roots from that of the control (Figs 7c,d, S5b). On the other hand, in transgenic bean hairy roots, even a significant malate exudation was detected under Al stress, but there were still significantly higher malate concentrations remaining in roots than those in the control (Fig. 8e,f). It has been well-documented that malate exudation plays important roles in Al detoxification of A. thaliana (Liu et al., 2012). However, citrate secretion has been reported to be a main Al tolerance mechanism in common bean (Stass et al., 2007; Rangel et al., 2010). Therefore, A. thaliana plants might have developed some quicker Al-responsive malate exudation pathways than bean, such as enhanced expression of AtAlMT1 (Hoekenga et al., 2006; Kobayashi et al., 2007). This is also partially proved by the fact that under Al stress, the malate exudation rates of WT A. thaliana plants were much higher than that of control lines of bean hairy roots (Figs 7d, 8f).

Organic acids in roots or in exudates play different roles in Al tolerance of plants. It has been characterized that external Al detoxification through malate exudation contributes greatly to Al tolerance in A. thaliana (Hoekenga et al., 2006; Liu et al., 2012). This is consistent with our results that the enhanced Al tolerance of overexpressing SgME1 A. thaliana lines was largely dependent on external Al detoxification through malate exudation (Fig. 7). However, besides the increased malate concentrations in both roots and exudates in overexpressing SgME1 bean hairy roots (Fig. 8), there was no significant change between control and OX lines in citrate accumulation or exudation (Fig. S6c,d). Thus, we summarize that Al tolerance in transgenic bean hairy roots combines an increase in internal Al detoxification through internal malate in roots with an accompanying malate exudation decreasing external Al toxicity.

We conclude that the superior Al tolerance of stylo genotype TPRC2001-1 is achieved mainly by internal Al detoxification through enhanced malate synthesis in roots and accompanied with Al exclusion via malate secretion. This process is functionally controlled by a novel malic enzyme, SgME1.


The authors acknowledge Dr Thomas Walk for corrections to the English. This research was jointly supported by the National Key Basic Research Special Funds of China (2011CB100301), the National Natural Science Foundation of China (grant no. 31025022) and the Earmarked Fund for China Agriculture Research System (CARS-35).